Table 1 in your link does not support 1 in a million for general anesthesia procedures, unless I’m badly misreading it.

Nor does the discussion section nor overall summary figure, which is 3 in a million for all forms of anesthesia (local, sedation, and general), where you have to at least exclude local anesthesia (the most common type) as having a fundamentally lower risk of death than sedation or GA.